Principal Investigators:
Karol Kowalski
Senior Research Scientist at Environmental Molecular Sciences Laboratory (EMSL) at Pacific Northwest National Lab
Co-PIs: Edoardo Apra, Eric Bylaska, Niranjan Govind, and Huub van Dam (Senior Research Scientists at EMSL)
Description:
The main goal of the Center is to modernize the NWChem suite of computational chemistry codes towards effective utilization of the emergent hybrid parallel computer systems based on the Intel Many Integrated Core Architecture (Intel MIC) technology. The proposed research presents a unique opportunity to make major breakthroughs in performance enhancements of several key implementations of many-body techniques which are indispensable for a comprehensive understanding of complex chemical transformations. Modernized codes will be applicable to several science drivers like, studies of aerosol particles, soil chemistry, biosystems, hormone-cofactor functionality in proteins, ionic liquids in cells, spectroscopies, new materials, and large-scale reaction mechanisms.
The emergence of heterogeneous systems will have a transformative impact on the landscape of high performance scientific computing. In particular, Intel MIC technology supports higher computational performance at lower cost and power consumption. We propose to develop new codes capable of taking advantage of these technological advancements.
A significant rewriting of algorithms is expected to take advantage of the new architecture. We plan to do these coding efforts in such a way that they will be easily available to integrate into other open-source computational chemistry software efforts. The extent to which code needs to be rewritten versus the extent to which code can be cleverly reused if the main challenges can be addressed or avoided depends to some degree on what the compilers for this platform afford. We will address the performance of several computational drivers in NWChem:
- Gaussian DFT and TDDFT methods
- Plane wave DFT formulations
- Multi-reference coupled cluster methods
We envision a major reworking of these codes with four major aims in mind: improvement of the parallelization up to hundreds of thousands of cores, modifications of the intra-node parallelization by using threading approaches, a complete rewrite of the major-kernels in order to better exploit the wide vector units available in the present and upcoming computer architectures, and reducing the computational cost for large scale ground and excited-state simulations.